$MRVL could benefit from both larger AI deployments and supplying more technology within each system. Custom processors, optical connectivity, switching and memory access give @MarvellTech several opportunities to expand its business with the same customers. Today’s expanded agreement with $GFS adds a manufacturing commitment that supports its optical growth plans. In my latest article, I examine how this connects with $GOOG, $NVDA and Celestial AI, why the long-term opportunity looks increasingly compelling, and which developments could turn that potential into higher cash flow per share. Full article 👇 Gracias, Pep. @damnang2 @PhotonCap @KawzInvests @michaelsikand @ren_stocks @CKCapitalxx @retail_mourinho

I’m really looking forward to $SIVE earnings report tomorrow. Beyond the headline numbers, reading between the lines will be crucial. I believe we could see at least one additional major deal announced. With production expected to begin in the first half of 2027, I’m also anticipating an update on the planned Nasdaq listing.

$NVTS still feels like one of the cleanest 800VDC plays out there. The company covers the whole power chain, from SiC on the grid side to GaN close to the GPU. Most people are focused on today’s small revenue base and completely miss the potential of the 2027 ramp. $NVTS helps power the entire AI buildout.

Did you know that $LITE and $VIAV were originally the exact same company? Today, $LITE is an independent corporation that develops optical components and lasers for data centers and artificial intelligence applications. However, before August 2015, the company did not exist on the stock market. It was part of a larger telecommunications conglomerate named JDS Uniphase (JDSU), which employed around 5,000 people globally. Wall Street investors pressured the management of JDS Uniphase to split the operation. The financial market preferred distinct companies focused on a single industry over a consolidated conglomerate that managed both hardware manufacturing and software testing. On August 1, 2015, the board finalized the division of the business into two independent entities: $LITE was formed by separating the optical components and commercial laser division. The new company started with 1,550 employees and an initial market capitalization of less than one billion dollars. $VIAV became the new name for the remaining core business of JDS Uniphase, keeping the network testing and software enablement divisions along with approximately 3,300 employees. Through this corporate action, the original joint operation was eliminated. $LITE took over the physical hardware manufacturing, while $VIAV retained the test and measurement infrastructure. Both companies now operate as separate publicly traded entities on the Nasdaq.

The probability that $AMD and $SIVE will publicly announce a commercial agreement over the coming quarters has increased significantly. While neither company has confirmed such a partnership, the available evidence points to a remarkable level of strategic alignment across manufacturing, supply chain integration, technology dependencies, and product roadmaps. Rather than relying on speculation or isolated rumors, this conclusion is based on multiple independent developments that all converge toward the same outcome. $SIVE has positioned itself at critical points of the silicon photonics ecosystem through partnerships with industry leaders, while $AMD has made substantial investments in optical interconnect technologies that require exactly the type of laser solutions $SIVE specializes in. When viewed together, these developments suggest that both companies are moving toward the same commercial destination. One of the strongest indicators is $SIVE deep integration into the manufacturing ecosystem already used by $AMD. $SIVE recently entered into a strategic collaboration with $GFS to integrate its indium phosphide (InP) distributed feedback (DFB) laser arrays directly into GF's silicon photonics platform. At the same time, $SIVE partnered with $JBL to manufacture 1.6-terabit optical transceivers aimed at eliminating bandwidth bottlenecks inside next-generation AI data centers. These partnerships are significant because $AMD does not manufacture optical components itself. Instead, it relies on an ecosystem of specialized manufacturing partners. $GFS is an established foundry partner for $AMD across multiple advanced semiconductor technologies, while $JBL serves as one of the industry's leading manufacturing partners for hyperscale infrastructure. As a result, $SIVE is no longer an outside supplier, it is becoming embedded within the same manufacturing ecosystem that supports $AMD future products. This dramatically lowers the barriers to adoption should $AMD decide to incorporate $SIVE laser technology into future optical architectures. $AMD is a strategic investor in @AyarLabs. $SIVE is the certified and exclusive supplier of the multi-wavelength external laser sources that power @AyarLabs SuperNova optical platform. This creates a direct technological dependency: If $AMD deploys @AyarLabs optical chiplet technology at scale, $SIVE laser solutions become an essential enabling component. $AMD further reinforced its commitment to silicon photonics through the acquisition of Enosemi, a company specializing in photonic integrated circuits for AI infrastructure. However, Enosemi does not manufacture compound semiconductor laser arrays. Scaling optical chiplet production therefore requires access to an experienced external supplier capable of delivering high-volume, high-reliability InP laser arrays. $SIVE is one of only a handful of independent companies worldwide with proven expertise in precisely this technology. The technological fit between $AMD optical ambitions and $SIVE manufacturing capabilities is therefore unusually strong. The expected commercial timelines of both companies also appear to converge. $SIVE has repeatedly communicated that customer demonstrations are nearing completion and that commercial production is expected to ramp throughout 2027. At the same time, $AMD next generation of AI accelerators is widely expected to introduce increasingly sophisticated optical interconnect technologies, including co-packaged optics (CPO) and optical chiplets, as traditional copper connections approach their physical limits. Commercial qualification for these platforms is expected during late 2026, followed by broader production in 2027. The synchronization of these timelines is difficult to ignore. Another potential catalyst is $SIVE planned U.S. NASDAQ dual listing. The company is currently transitioning its financial reporting to PCAOB standards in preparation for listing in the United States. During this process, companies are generally required to disclose material commercial relationships and significant customers when those relationships are considered material to investors under applicable SEC disclosure rules. With $SIVE reporting a commercial pipeline approaching USD 799 million, any major agreement contributing materially to that pipeline could become subject to disclosure requirements if finalized. While the exact disclosure obligations depend on the facts and applicable SEC rules, the listing process could increase the likelihood that previously undisclosed commercial relationships become public. Recent insider activity provides an additional, although indirect, indicator. Following the expiration of trading restrictions, the CEO and several members of $SIVE Board of Directors executed coordinated open-market purchases of company shares.